Objective To evaluate the influence of dynamic fixation (rotating and sliding pedicle screws) on stability of the atlantoaxial joint. Methods A series of in vitro biomechanical tests were performed using six fresh adult cervical spines (occipital bone-C4 segment) to simulate different conditions in surgery， including the intact state, the injury state, rigid fixation, rotating pedicle screw fixation, sliding pedicle screw fixation. The repeated measurement design was employed, and under intact, injury and different fixation states, the pure moment of 1.5 N·m in flexion-extension, left-right lateral bending, left-right axial rotation directions were applied using the spinal testing machine. The movement of atlantoaxial spine was measured consecutively by three-dimensional （3D） measurement system in order to analyze the range of motion (ROM) and neutral zone (NZ) of atlantoaxial joints. Results Under injury state, ROM of atlantoaxial joints was significantly larger than that under intact state during flexion, extension, lateral bending and rotation, leading to the instability of atlantoaxial joints. ROM of fixation segments was significantly reduced during flexion, extension, lateral bending and rotation after rigid and dynamic fixation. Compared with rigid fixation, dynamic fixation showed a significant ROM increase during lateral bending. NZs of fixation segments after dynamic fixation were significantly reduced. There were no significant ROM differences between rigid fixation and dynamic fixation. Conclusions The stability of atlantoaxial joints by dynamic fixation during flexion, extension and rotation was comparable to that by rigid fixation, but weaker during lateral bending. Dynamic screw fixation can maintain the relative stability of atlantoaxial joints.